1) The trades are showing entries and exits on days the market is closed (Saturdays and holidays). All the trade dates appear to be one day later than the actual execution. So for example, if the trade is supposed to execute on the open of a Friday, it executes at that Friday's opening price (more or less), but the trade date shown is the next day, i.e., Saturday. This doesn't seem to be affecting the results, but it's doesn't build confidence either, and it's a pain when creating formulas in excel that need to refer to the date.
2) A subset of the trades show execution prices that are way out of whack, i.e., more than 3% different (many 6.5% to 11.9% different) than they should be, even assuming built-in slippage. Of the 1489 trades executed in the backtest (purchase + sale = 1 trade), 157 exhibit this problem.
Despite this, the entries and exits appear to be correct with regard to timing, but it means the results of the backtest aren't correct.
So my questions are these. Is problem #1 a known bug? Is there a fix for this? Does it matter? And for question #2, is this also a known bug? Is there something I'm doing wrong? I'll paste the key logic of the backtest code below. (can't paste it all due to length restrictions)
Thanks for looking at this,
Kim G
Comment